Researchers Flip the Switch on the World's Largest Neuromorphic Supercomputer

Reviving the Supercomputer

For most of us, when we hear the term supercomputer, images of a grossly oversized, distant relative of the modern day laptop come to mind. This one, however, which the researchers have named SpiNNaker, or Spiking Neural Network Architecture, is directly modeled after the human brain. It's known as a neuromorphic supercomputer, and reflecting the symbolism of the event, the machine received its one-millionth processor core.

After 20 years in conception, over 10 years in construction & £15 million in funding, today the SpiNNaker machine reaches its 1,000,000 cores milestone @FurberSteve @apt_csmcr @csmcr @HumanBrainProj . Read our full story #CodingTheFuture #supercomputers https://t.co/NWMkSVrQ4A pic.twitter.com/XUIN3t6glo

— Manchester Uni News (@UoMNews) November 2, 2018

In dramatic form, it was switched on for the first time on Friday. It is being used to control a robot named, and in terms of what makes the supercomputer capable of running more processes based on biological neurons--in real time--than any other machine of its kind in the world:

--> Each chip possesses 100 million moving parts.

--> It can carry out more than 200 million actions

A Landmark Contribution in the Growing Field of Neuromorphic Computing

The field of neuromorphic computing essentially involves using machines to indirectly take a glimpse inside the human brain. Initiatives like the Human Brain Project, out of which the SpiNNaker project was conceived, is helping provide more clues. In order for work in this important field to flourish, however, what is suggested by researchers is "a paradigm shift for computing as a whole" as well as the development of a "completely new category of hardware".

SpiNNaker stands out for a few important reasons, but most significantly because of its processing capacity. Researchers hope one day to up the ante even more and fit the supercomputer with its one billionth processor, which would add up to the power of 1% of the neurological capacity of the human brain--the organ which possesses about 100 million neurons in total.

Steve Furber, Professor of Computer Engineering, who first imagined a supercomputer of this scale, said of the grand mission of the project: “SpiNNaker completely re-thinks the way conventional computers work. We’ve essentially created a machine that works more like a brain than a traditional computer, which is extremely exciting. The ultimate objective for the project has always been a million cores in a single computer for real-time brain modeling applications, and we have now achieved it, which is fantastic.”

A Far-reaching Impact

The successful execution of SpiNNaker has two important benefits: (1) the neuroscientist community stands to deepen its understanding of brain function, and (2) the robotics industry can employ the "brain power" needed for controlling robots.

Professor Furber summed up the tremendous advantages that a neuron-modeled machine can have in the realm of biology: “Neuroscientists can now use SpiNNaker to help unlock some of the secrets of how the human brain works by running unprecedentedly large-scale simulations. It also works as a real-time neural simulator that allows roboticists to design large-scale neural networks into mobile robots so they can walk, talk and move with flexibility and low power.”

For SpOminbot, a robotic platform created by a team of engineers at Technische Universität München, which is fitted with three omniwheels and two silicon retinas, SpiNNaker provided valuable assistance to the bot--via a CPLD and ARM microcontroller--in terms of helping it to assess the environment and then navigate around potential obstacles.